Search

The vertical structure of wind is a key factor in modulating air quality, from which the determination of boundary layer height (BLH) remains a major challenge. In this paper, we developed an improved threshold method to determine the BLH from radar wind profiler (RWP) measurements. The normalized signal-to-noise ratio (SNR) profiles were used instead of the original SNR profiles to avoid instrumental inconsistencies. Additionally, a peak filter was designed to indicate the BLH based on the maximum SNR by taking into account the multiple peaks in the SNR profile. This algorithm was then applied to the RWP measurements taken in the summer (June–July–August) of 2018 in Beijing to obtain the BLHs. Validation analyses suggested that the BLH retrievals from RWP exhibited high consistency with those from radiosondes, with an average correlation coefficient of 0.69 (0.66) and a root mean squared error of 0.39 (0.41) in the daytime (nighttime). Additionally, the major features of summertime BLHs in Beijing were examined. In particular, a distinct diurnal variation in BLH was observed with a peak (1630 ± 510 m) occurring at 0600 universal time coordinated (UTC) and a minimum (587 ± 343 m) at 2300 UTC. Therefore, the algorithm presented here has great potential to be applied to other regions to obtain reliable BLHs. The findings obtained here highlight the importance of vertical wind structure in air quality studies. [ABSTRACT FROM AUTHOR]

The interpolation, prediction, and feature analysis of fine-gained air quality are three important topics in the area of urban air computing. The solutions to these topics can provide extremely useful information to support air pollution control, and consequently generate great societal and technical impacts. Most of the existing work solves the three problems separately by different models. In this paper, we propose a general and effective approach to solve the three problems in one model called the Deep Air Learning (DAL). The main idea of DAL lies in embedding feature selection and semi-supervised learning in different layers of the deep learning network. The proposed approach utilizes the information pertaining to the unlabeled spatio-temporal data to improve the performance of the interpolation and the prediction, and performs feature selection and association analysis to reveal the main relevant features to the variation of the air quality. We evaluate our approach with extensive experiments based on real data sources obtained in Beijing, China. Experiments show that DAL is superior to the peer models from the recent literature when solving the topics of interpolation, prediction, and feature analysis of fine-gained air quality. [ABSTRACT FROM AUTHOR]